X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Fextractor%2Fappletrailers.py;h=87580147dcb4176b163417053ac5a2489cf556c3;hb=9e1a5b845586a0a5431fb72467142046d8571e6f;hp=4359b88d1b7057944beb126eb8a1c82dbb818758;hpb=1c1cff6a525bc8fc506cf2c6eb8963abc3b1fcee;p=youtube-dl diff --git a/youtube_dl/extractor/appletrailers.py b/youtube_dl/extractor/appletrailers.py index 4359b88d1..87580147d 100644 --- a/youtube_dl/extractor/appletrailers.py +++ b/youtube_dl/extractor/appletrailers.py @@ -70,11 +70,13 @@ class AppleTrailersIE(InfoExtractor): uploader_id = mobj.group('company') playlist_url = compat_urlparse.urljoin(url, 'includes/playlists/itunes.inc') + def fix_html(s): s = re.sub(r'(?s).*?', '', s) s = re.sub(r'', r'', s) # The ' in the onClick attributes are not escaped, it couldn't be parsed # like: http://trailers.apple.com/trailers/wb/gravity/ + def _clean_json(m): return 'iTunes.playURL(%s);' % m.group(1).replace('\'', ''') s = re.sub(self._JSON_RE, _clean_json, s) @@ -86,7 +88,7 @@ class AppleTrailersIE(InfoExtractor): for li in doc.findall('./div/ul/li'): on_click = li.find('.//a').attrib['onClick'] trailer_info_json = self._search_regex(self._JSON_RE, - on_click, 'trailer info') + on_click, 'trailer info') trailer_info = json.loads(trailer_info_json) title = trailer_info['title'] video_id = movie + '-' + re.sub(r'[^a-zA-Z0-9]', '', title).lower()